1211932878 Q * dowdle Remote host closed the connection 1211934163 Q * ViRUS Quit: Leaving 1211935650 J * jenny20 ~jenny20@d033.dhcp212-198-248.noos.fr 1211935767 Q * jenny20 1211936725 J * doener ~doener@i577BAB6E.versanet.de 1211936829 Q * doener_ Ping timeout: 480 seconds 1211940926 J * NetNuttt ~NetNuttt@adsl-065-006-153-049.sip.asm.bellsouth.net 1211942485 Q * NetNuttt Remote host closed the connection 1211945755 M * Bertl okay, off to bed now .. have a good one everyone! 1211945761 N * Bertl Bertl_zZ 1211946619 Q * pflanze Ping timeout: 480 seconds 1211947673 Q * lagann Quit: Leaving 1211947812 J * lagann ~distro@c-66-30-110-51.hsd1.ma.comcast.net 1211948894 J * cryptronic ~oli@p54A3B58A.dip0.t-ipconnect.de 1211950811 Q * cryptronic Quit: Leaving. 1211951931 J * sharkjaw ~gab@64.28.12.166 1211952000 J * Slydder ~chuck@194.59.17.53 1211954326 J * bfremon ~ben@lns-bzn-20-82-64-27-28.adsl.proxad.net 1211954910 Q * doener Ping timeout: 480 seconds 1211956224 J * rgl ~rgl@bl8-133-78.dsl.telepac.pt 1211957497 M * rgl Bertl_zZ, are you using or following openvz? 1211957624 Q * meandtheshell Quit: Leaving. 1211958740 J * dna ~dna@149-195-dsl.kielnet.net 1211958880 M * arekm hello, did anyone try to install cpanel in vserver? I wonder if it needs things that vserver doesn't allow to access from guest (like changing network/iptables and so on) 1211958906 M * arekm (current setup is xen + centos which I would like to throw away and use vserver of course) 1211959516 N * DoberMann[ZZZzzz] DoberMann 1211959537 J * meandtheshell ~sa@d91-129-52-22.cust.tele2.at 1211959706 Q * balbir Ping timeout: 480 seconds 1211959830 M * rgl arekm, why are you ditching xen? 1211960030 Q * xdr Ping timeout: 480 seconds 1211960170 M * arekm rgl: too complex for very little gain in my setup comparing to vserver 1211960434 M * rgl arekm, what you mean by complex? can you be more specific :) 1211960454 M * rgl arekm, running multiple kernels? 1211960542 M * arekm rgl: voodoo magic with network scripts on dom0 for example. When it works then there is no problem. When it doesn't work then you can spend few hours digging into xenisms ;> 1211960629 M * arekm I like shared disk space in vserver, too. possible in xen with nfs only afaik 1211960667 J * yarihm ~yarihm@84-74-147-84.dclient.hispeed.ch 1211960860 J * ktwilight_ ~ktwilight@95.208-66-87.adsl-static.isp.belgacom.be 1211960885 M * rgl arekm, humm I've used the bridging and in fact the script are somewhat exotic, though, I think the diteched them in the current tree. 1211960927 M * rgl arekm, you can do shared with vserver? 1211960974 M * rgl arekm, the only "shared" stuff I known about is the COW stuff that is made between the guests, but that stuff is not visible to the guests. 1211960983 M * arekm that's the other problem. xen is changing quite fast while docs are not... so you have tons of very different docs about the same area (xencons= option for example) 1211960995 M * arekm rgl: shared == single filesystem shared among few guests 1211961056 M * rgl arekm, but that sharing is somewhat "read-only", I mean, its copy-on-write. 1211961097 Q * ktwilight Read error: Connection reset by peer 1211961122 M * arekm rgl: ah, I'm not talking about sharing files. Just sharing single partition so free space is one piece for all guests 1211961144 M * rgl arekm, I mean, I only see that as performance advantage. do you use for other purpose? 1211961156 J * ktwilight ~ktwilight@181.208-66-87.adsl-static.isp.belgacom.be 1211961223 M * arekm rgl: in xen 100GB disk shared among 5 guests need to be split for example 20GB for each guest. Free space on 20GB guest ends and you have a problem. On vserver each guest sees 100GB of space (that can be consumed by any of the guests) 1211961237 M * arekm (s/on vserver/on one of possible vserver setups/) 1211961309 M * rgl I didn't known you could do that *G* 1211961326 M * rgl but I don't quite like that anyways hehehe 1211961352 M * rgl its a provisioning problem, supposed all the guests really want 100G? :D 1211961377 Q * ktwilight_ Ping timeout: 480 seconds 1211961380 M * rgl (for my use case that is ;) 1211961395 M * arekm you can't win in such case, not with vserver and not with xen unless you add more disks 1211961415 M * mugwump vserver has disklimits for that 1211961421 M * mugwump you limit each server to, say, 40GB 1211961429 M * rgl indeed. in my use case, I would have to add more disks. 1211961451 M * arekm mugwump: afaik that functionality is very limited, some fses only afaik (I can be wrong :) 1211961468 M * rgl I use lvm. if I need to grow the space, I just resize the partition. 1211961493 M * arekm rgl: what fs? 1211961520 M * rgl arekm, plain old ext3. 1211961529 M * arekm the problem is shrinking one to be able to grow another one under some fses. 1211961544 M * rgl humm, never needed to shrink :D 1211961576 M * rgl I guess I would have to tar it up, and recreate a volume. 1211961645 M * rgl have you compared performance of xen vs vserver? 1211961672 M * arekm no, even +-10% won't make any difference for me 1211962083 J * balbir ~balbir@59.145.136.1 1211962216 M * mugwump it works on ext3, what more do you need? :) 1211962230 M * mugwump there's a nice paper comparing xen vs vserver performance on the wiki 1211962249 M * arekm mugwump: xfs 1211962291 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1211962407 M * mugwump xfs is good for certain very domain specific uses, for everything else it sucks IME 1211962424 M * mugwump throw any significant number of inodes at it and it falls over 1211962443 M * mugwump "crap, why is my checkout taking so long, is something wrong? Oh, yeah, xfs" 1211962864 M * arekm vserver centos5 build --force -n centos5 --context 1400 -m yum -- -d centos5, hm, is it possible to pass some options like "-v" to yum? 1211962941 M * arekm beside patching vyum-worker 1211963076 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1211963226 M * Hawq I tried to shrink ext3 once, worked ok 1211963307 M * Hawq so with lvm+ext3 I can resize whatever I want, theoretically 1211963323 Q * yarihm Quit: This computer has gone to sleep 1211963338 J * ISSAMNEO1 ~ISSAMNEO1@213.150.170.98 1211963346 M * ISSAMNEO1 Good morning 1211963415 M * ISSAMNEO1 i create a vserver using rsync from an FAI (dhcp + NFS) 1211963433 M * ISSAMNEO1 on my vserver dhcp work but not NFS 1211963624 J * jmcarica` jm@183.204.195-77.rev.gaoland.net 1211963651 Q * jmcarica` Remote host closed the connection 1211963802 M * ISSAMNEO1 i try to start it manually (/etc/init.d/dhcp-kernel-server start) but it shows can't load module in /lib/modules/kernel version of the host/modules.dep, and since the kernel version of my host and the kernel version of my guest are not the same, so it fail 1211963825 M * arekm hm, newer yums hang at importing gpg key when creating centos guest 1211963932 M * arekm has anyone seen that? 1211963992 M * franck34 arekm: yep, but nothing related to vserver, make a "top" a look at the %CPU yums is using 1211964025 M * franck34 hi all, is somebody get experiences using a box with 2 vserver inside: one for mailserver, one for webserver ? 1211964054 M * franck34 the problem is that the host and the 2 vserver need to send mails, and the mailserver can receive mails 1211964063 M * franck34 so port 25 is already binded on the host ... 1211964090 M * franck34 i just started with vserver, perhaps i choose a bad strategy 1211964101 M * franck34 my goal was to be able to switch postfix/qmail just changing an iptable rules 1211964115 M * franck34 but seem's i'll have problem because of bind port 25 every where 1211964129 M * franck34 any idea/suggestion/critism/idea are welcome thanks in advance 1211964236 M * PowerKe how many instances of postfix/qmail are you trying to run? 1211964293 M * PowerKe I just put postfix in 1 guest and use that for all mail needs. The others just have a simple mailscript that will deliver the mail to the postfix guest which will take care of getting the email delivered 1211964302 M * franck34 1 vserver using qmail, one other using postfix. But when i speak of "switch" is stop qmail then start postfix, and vice et versa 1211964314 M * franck34 PowerKe: ha, 1211964322 M * franck34 PowerKe: i think i'm looking for that 1211964329 M * PowerKe stopping one, starting the other should work as well 1211964334 M * franck34 PowerKe: yep 1211964352 M * franck34 PowerKe: what mailscript are you using ? (php is looking for /usr/sbin/sendmail by example ...) 1211964390 M * PowerKe mail-mta/ssmtp package (Gentoo) 1211964404 M * franck34 i'm on debian let's apt-cache search 1211964405 M * franck34 thx 1211964432 M * franck34 ssmtp - extremely simple MTA to get mail off the system to a mail hub 1211964509 M * PowerKe either that or use a tcp socket connection to the postfix guest for apps that support it 1211964527 M * arekm franck34: 0.4% 1211964590 M * franck34 PowerKe: i was thinking of that too, but seem's ssmtp is what i was looking for thanks a lot. Now i just have to work on the conf (relay, ...) 1211964593 M * franck34 arekm: ? 1211964602 J * yarihm ~yarihm@mtec-hg-docking-1-dhcp-6.ethz.ch 1211964620 M * arekm franck34: it doesn't eat cpu. It's doing select() waiting for something to happen... 1211964641 M * PowerKe franck34: If you're binding ports on the host, make sure to restrict them to the host IP, that should avoid problems when trying to bind to the same port in a guest (on a different ip) 1211965312 J * Sniper ~snipper.f@222.66.69.238 1211965503 M * franck34 PowerKe: yeah, i was looking for that too, but i'm familiar with qmail only, don't know where to setup bind ip for postfix 1211965518 M * mugwump ISSAMNEO1: you might need to add network caps 1211965521 M * franck34 what's why i want 2 vserver, one for qmail, the other one with postfix: to learn postfix ;) 1211965532 M * franck34 arekm: i have no idea sry 1211965555 M * franck34 arekm: just wait if it's the first time you are launching yum 1211965561 M * mugwump ISSAMNEO1: or disabling portmap on the host etc 1211965596 M * mugwump actually, all I needed for a FAI vserver here was secure_mount and mknod caps, for running fai-setup (and hence mkfainfsroot) 1211965814 M * PowerKe franck34: If you put both mailservers in different guests, you shouldn't have bind problems (guests can only bind to their own ip's) 1211965877 Q * Sniper 1211966098 M * franck34 PowerKe: agree with that got a bad conf at the moment cause i already have smtp binding on the host 1211966126 M * franck34 so i need to use ssmtp on the host to relay mail into my mail vserver 1211966137 M * franck34 and setup my web vserver to use ssmtp too 1211966162 M * franck34 PowerKe: btw, what are you using for antispam in postfix ? 1211966248 M * ISSAMNEO1 mugwmp : i have an fai server working fine 1211966268 M * ISSAMNEO1 and we want now to add it as a virtual server 1211966309 M * ISSAMNEO1 so i build a virtual server using the rsync method (copy my old server to the new vserver) 1211966344 M * ISSAMNEO1 after i finish the build i start the virtual server but as i say i find that nfs isn't working 1211966436 M * franck34 when i ping google, first vserver show 10ms, but on the second in the same time, 30ms and seem's it's lagging 1211966448 M * PowerKe franck34: amavis (wrapper script that calls spamassassin and antivirus) 1211966469 M * franck34 PowerKe: k thx 1211966491 M * ISSAMNEO1 for the status of nfs is: 1211966515 M * ISSAMNEO1 nfs-kernel version ==> nfsd not running 1211966553 M * ISSAMNEO1 nfs-common ==> rpc.statd runing, but rpc.ipmapd halted 1211966653 J * Sniper ~snipper.f@222.66.69.238 1211968461 J * bfremon1 ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1211968760 Q * bfremon Ping timeout: 480 seconds 1211970608 N * Bertl_zZ Bertl 1211970615 M * Bertl morning folks! 1211970817 Q * jsambrook Remote host closed the connection 1211970896 M * hijacker morning Bertl 1211971078 J * jsambrook ~jsambrook@aelfric.plus.com 1211971262 J * pflanze ~chris__@77-56-83-98.dclient.hispeed.ch 1211971297 Q * Sniper 1211972672 Q * ISSAMNEO1 Ping timeout: 480 seconds 1211974146 Q * sharkjaw Ping timeout: 480 seconds 1211974153 J * ISSAMNEO1 ~ISSAMNEO1@213.150.170.98 1211974454 J * sharkjaw ~gab@64.28.12.166 1211974661 Q * yarihm Quit: This computer has gone to sleep 1211974791 Q * bfremon1 Quit: Leaving. 1211974965 J * bfremon ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1211977423 Q * balbir Ping timeout: 480 seconds 1211977748 J * yarihm ~yarihm@vpn-global-dhcp2-79.ethz.ch 1211978219 J * balbir ~balbir@59.145.136.1 1211978303 Q * yarihm Quit: This computer has gone to sleep 1211978627 Q * Aiken Remote host closed the connection 1211980502 Q * sharkjaw Remote host closed the connection 1211982109 J * Surfer47 ~javachat@cpe-74-71-143-65.twcny.res.rr.com 1211982109 A * Surfer47 Visit http://www.FakeMagazineCover.com (upload pic make mag) - http://www.SillyWebcam.com (play with webcam online) - http://www.Is-A-Jerk.com (insulter/anon email) - http://www.ComedySearchEngine.com (fun) - http://www.BodySwitcher.com (put your face on funny body) - http://www.MedChecker.com (health) - http://www.Canuckster.com (Canada eh) - http://www.Nerdful.com (geeks) 1211982114 Q * Surfer47 1211982153 M * jesusch just bbullshit 1211983081 Q * Slydder Quit: Leaving. 1211983264 J * doener ~doener@i577BAB6E.versanet.de 1211983753 M * ISSAMNEO1 no one to help 1211983754 M * ISSAMNEO1 !!! 1211983826 M * ISSAMNEO1 i create a vserver using rsync my old server and the new server don't have the same kernel 1211983833 M * ISSAMNEO1 right 1211983861 M * Bertl tht isn't a problem 1211983864 M * Bertl *that 1211983878 M * Bertl just make sure that you do the rsync properly 1211983932 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it 1211983952 M * ISSAMNEO1 but modules.dep in my old server and on the vserver are not in the same place since 1211983961 M * ISSAMNEO1 the modules.dep location is 1211983993 M * Bertl you don't need modules.dep inside a Linux-VServer guest, and the hostshould be fine as it is 1211984063 Q * mrfree 1211984077 M * ISSAMNEO1 so why when i try /etc/init.d/nfs-kernel-source start on guest it don't work and give an error : can't load modules.dep 1211984114 M * Bertl because you cannot run an nfs kernel server inside a Linux-VServer guest 1211984133 M * Bertl i.e. it doesn't make sense to try to load a module there in the first place 1211984245 M * ISSAMNEO1 and when my guest need nfs how i do it, i think i install it on the host then what? 1211984272 M * Bertl depends on what your guest actually needs .. 1211984287 M * Bertl do you want to mount an nfs volume inside a guest? 1211984299 M * Bertl do you want to export a directory inside a guest? 1211984373 M * ISSAMNEO1 no the guest will be an FAI server (DHCP and NFS server) 1211984386 M * ISSAMNEO1 nfs volume 1211984444 M * Bertl well, not the best choice for a Linux-VServer guest 1211984475 M * Bertl but you want to export some guest subdirectory to the world then 1211984489 M * Bertl (besides running dhcpd inside the guest) 1211984603 M * ISSAMNEO1 not problem the FAI server is supposed install linux on client pc so it must contain an NFS server , and client will download the installation from it 1211984624 M * Bertl yes, I understand that 1211984660 M * Bertl but if you want to put the nfs server _on_ the guest (isntead of using the host nfs server) you have to use some userspace nfs implementation, not the kernel nfs 1211984831 M * micah i managed to rsync a vserver to a backup server without the right options, so when I restored it, I dont have the necessary device files 1211984851 M * micah so when I try to enter the guest, I get: vlogin: openpty(): No such file or directory 1211984861 M * micah is there an easy way to re-create those? 1211984877 M * Bertl yep, you create a skeleton guest and copy them from there 1211984895 M * micah probably just things from /dev ? 1211984896 M * Bertl (cp -va will do the trick) 1211984912 M * Bertl all devs are in dev, so if your problems are 'devices' then yes 1211984927 M * micah i hope thats it 1211984931 M * Bertl note that you probably want to restart the guest 1211984950 M * micah i think the rsync was done with -a -v --numeric-ides 1211984952 M * Bertl if you still have the original, you can re-sync to make sure (With the proper options) 1211984957 M * micah --numeric-ids 1211985026 M * Bertl for the record, good rsync options are: 1211985049 M * Bertl -axHPSD --numeric-ids (add -c if you want to checksum, -z if you want to compress) 1211985065 M * micah thanks 1211985101 M * Bertl you're welcome! 1211985187 M * ISSAMNEO1 i have 2 dir in my guest to share them through nfs what is the best solution 1211985203 M * Bertl simply export them on the host 1211985230 M * micah looks like copying the /devices from the skeleton worked! 1211985308 M * ISSAMNEO1 and how the client will see them since the host and the guest have different @ IP 1211985346 M * micah although I think I have some permissions to fix 1211985366 M * Bertl ISSAMNEO1: the host always has _all_ IPs, including that of the guests 1211985414 M * Bertl *those 1211985420 Q * bfremon Read error: Connection reset by peer 1211985464 M * micah uff, yeah all file/directory ownerships are messed up 1211985651 J * bfremon ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1211986011 M * rgl OT: you guys known how can I set the supplementary group IDs of the running process? (using C/python/whatever ;) 1211986081 M * daniel_hozac setgroups 1211986612 M * franck34 me again, i've just create a newvserver, but got /tmp really small, how can i solve that ? 1211986662 M * Bertl newvserver or new vserver? (if the former, don't do that :) 1211986712 M * daniel_hozac edit /etc/vservers//fstab 1211986723 M * Bertl tmp is intentionally small (16M) by default, your guest should use /var/tmp for larger storage, but if you really want to make /tmp larger (it's a tmpfs) you can adjust it in /etc/vservers//fstab, or completely remove it 1211986758 M * franck34 thanks a lot 1211986791 M * franck34 let's try /var/tmp first 1211986840 J * dowdle ~dowdle@scott.coe.montana.edu 1211986892 Q * FireEgl Quit: Leaving... 1211987709 M * franck34 can /tmp can be a symb link to /var/tmp ? 1211987784 M * Bertl sure, but if you just want to get rid of /tmp (as tmpfs) it's better to remove the mount from the fstab 1211987811 M * franck34 in fact, nothing speaking of /tmp in my fstab ... 1211987816 M * franck34 but df show 17Mo 1211987865 M * daniel_hozac then you're not looking at the right fstab. 1211987896 M * Bertl we are talking about /etc/vservers//fstab :) 1211987931 Q * balbir Ping timeout: 480 seconds 1211988053 M * transacid hmm /etc/init.d/ssh: line 78: /proc/21272/oom_adj: Operation not permitted 1211988069 M * transacid i suddenly get that in every vserver guest 1211988121 M * Bertl guess your 'new' sshd is trying to mess with the oom adjustments 1211988139 M * Bertl just remove that line (or ignore if it works anyways) 1211988375 M * daniel_hozac i don't see what returns EPERM though. 1211988464 M * transacid Bertl: commenting the ifclause out helped 1211988472 M * transacid dunno what it's for 1211988482 M * transacid i hope not essentiel :) 1211988529 M * Bertl it makes sure, that the oom killer doesn't reap your sshd 1211988547 M * Bertl (rendering the host unusual, because inaccessible) 1211988599 M * franck34 in a vserver, i'm trying to make an ssh jail 1211988610 M * franck34 while trying to recreate /dev in the jail got this error 1211988612 M * franck34 mknod -m 0666 /home/chrooted/dev/null c 1 3 1211988619 M * franck34 operation not permitted 1211988623 M * franck34 something related to vserver ? 1211988630 M * Bertl yep, device creation is forbidden inside guest 1211988641 M * Bertl (otherwise you could easily mess up the host) 1211988649 M * franck34 ok so .. 1211988656 M * franck34 not possible to make a jail ? 1211988659 M * Bertl create them on the host, or copy them there 1211988675 M * pflanze hardlinking could work 1211988675 M * franck34 but /dev is existing on the host 1211988677 M * daniel_hozac if you have 2.3, you can set it up to allow that too. 1211988699 M * franck34 pflanze: an example plz ? 1211988707 M * franck34 just ln without "-s" ? 1211988710 M * franck34 daniel_hozac: wow 1211988712 M * pflanze exactly 1211988726 M * Bertl franck34: the problem is, if you allow the guest to create arbitrary devices, then that could easily create a device for your harddisk, and fill it with zeroes 1211988737 M * franck34 Bertl: i completely understand 1211988762 M * franck34 pflanze: ln /data/vserver/myvserver/dev /dev 1211988767 M * franck34 does it make sense ? 1211988773 M * daniel_hozac no, you can't hardlink directories. 1211988774 M * pflanze franck34: no I meant from inside the vserver 1211988782 M * pflanze and the individual device files, yeah 1211988798 M * mnemoc bind them 1211988812 M * Bertl mnemoc: lotr? 1211988825 M * franck34 wowow guys, first time i receive so many suggestion on an irc chat from different people :) 1211988831 M * franck34 cool 1211988854 M * franck34 pflanze: so ln /home/chrooted/dev /dev inside my vserver ? 1211988870 M * Bertl franck34: 'man ln' :) 1211988873 M * pflanze franck34: no, ln /dev/* /home/chrooted/dev/ 1211988879 M * franck34 haa thxc 1211988889 M * pflanze and look out for subdirs 1211988894 M * mnemoc Bertl: eh? o.o 1211988911 M * pflanze only works if both dirs are on the same filesystem, of course 1211988925 M * franck34 it suppose to be 1211988937 M * Bertl mnemoc: ... and in the darkness 'bind them' :) 1211988957 M * franck34 doesn't work 1211988968 M * franck34 "direct link is not permitted for a directory" 1211988980 M * pflanze I said, look out for subdirectories 1211988988 M * franck34 and if i try one by one, got "crossed link of invalid device" 1211988989 M * Bertl like /dev/pts 1211989007 M * franck34 k 1211989013 M * franck34 just need /dev/null in fact 1211989025 M * mnemoc Bertl: :D 1211989039 M * franck34 working 1211989040 M * franck34 testing 1211989058 M * franck34 yeah you guys are all genius here 1211989060 M * franck34 thanks a lot 1211989075 M * franck34 an sftp jail in a vserver ... 1211989080 A * franck34 happy 1211989085 M * Bertl congrats! 1211989146 M * pflanze cp -rl /dev/. /home/chrooted/dev/. will do as much as it can 1211989167 M * franck34 pflanze: thanks a lot i'm putting all that in my notes ;) 1211989236 M * franck34 finishing a qmailrocks installation in a vserver, if this interesting somebody here 1211989257 M * Bertl would be nice to add some of those notes to the wiki (see topic :) 1211989259 A * franck34 running out because afraid of be killed because of using "qmail" word 1211989282 M * Bertl nah, we all deal fine with different universes (like the djb one) 1211989300 M * franck34 hehe 1211989315 M * pflanze I'm using sort of a distributed qmailrocks install: 1211989329 M * franck34 website is speaking about old/new wiki 1211989330 M * pflanze clamav in one vserver, qpsmtpd in another, more qmail in yet others. 1211989334 M * franck34 what is the good one ? 1211989347 M * franck34 pflanze: wow nice 1211989537 M * pflanze the new wiki is the right one, I guess 1211989852 M * franck34 ha ok the current one in linux-vserver.org is the good one 1211989871 M * franck34 i notice 2 things today if there nothing speaking of that in the wiki 1211989883 M * franck34 story of initab for qmailrocks 1211989893 M * franck34 and /dev creating for jail 1211990103 M * franck34 in the FAQ ? 1211990161 M * franck34 or must a create 2 new page something like SSH Jail in vserver and one other is qmailrocks installation in a debian vserver ? 1211990231 M * Bertl well, first, I'd try to make it as distro agnostic as possible 1211990254 M * franck34 like you want 1211990261 M * franck34 is the story of inittab particular to debian ? 1211990282 M * Bertl (we don't want pages like, how to setup ftp in debian, redhat, fedora, gentoo ...) 1211990290 M * franck34 i see 1211990307 M * Bertl but, we want pages like: how to fix the broken ftp setup in ubuntu :) 1211990338 M * Bertl i.e. makes sense to add a distro specific hint/workaround, but generalize on common things 1211990350 M * franck34 i'd like to say something like 'add "plain" into /etc/vservers/XXX/apps/init/style to allow initab to be taken at vserver "boot" time, usefull for qmailrocks by example' 1211990368 M * franck34 so should be 1211990376 M * daniel_hozac http://linux-vserver.org/util-vserver:InitStyles 1211990395 M * franck34 "how to make svscan starting a vserver boot time" 1211990421 M * franck34 s/a/at 1211990524 M * Bertl what's svscan? 1211990534 M * franck34 http://linux-vserver.org/Howto_launch_Svscanboot_At_Boot_Time 1211990538 M * franck34 here ? 1211990552 M * franck34 it's something i have in my /etc/inittab 1211990561 M * franck34 SV:123456:respawn:/command/svscanboot 1211990564 M * daniel_hozac is it guest-specific? 1211990566 M * Bertl what does it do? 1211990578 M * daniel_hozac i.e. is that not something you need to do if you setup qmail on a regular box? 1211990607 M * franck34 Bertl: it permet to a daemon to scan if all qmail services are up and running, trying to restart them if not 1211990618 M * franck34 daniel_hozac: yes, problem it's in the qmailrocks installation guide 1211990638 M * franck34 if you google, you will find only an IRC log telling this story of "plain" in init/style ;) 1211990650 M * Bertl franck34: hmm, isn't djb's universe driven by daemontools? 1211990662 M * franck34 Bertl: probably, i'm not an expert 1211990690 M * Bertl I mean, looks like you are adding an inittab entry for sysv init? 1211990697 M * franck34 just telling you a point of view of a noob (me) 1211990699 M * franck34 yes it is 1211990722 M * franck34 ha you mean Sxx, Kxx ? 1211990725 M * franck34 init.d ? 1211990750 M * pmenier hello 1211990774 M * pmenier franck34: you could use sys_alive to scan qmail services 1211990777 M * daniel_hozac franck34: see the link i pasted. it's got descriptions of the init styles. 1211990795 M * franck34 pmenier: probably 1211990802 Q * ISSAMNEO1 1211990803 M * Bertl franck34: what is your guest distro? 1211990809 M * franck34 Bertl: debian 1211990811 M * franck34 4.0 1211990833 M * Bertl so you probably want to use sysv init there (unless you want a separate init hanging around) 1211990840 M * franck34 for all people just notice i don't have any problem, just trying to put a usefull entry in the wiki to avoid loosing time googling for next people ;) 1211990873 M * Bertl what is your init style atm= 1211990877 M * Bertl s/=/? 1211990877 M * rgl OT: I'm having a bit trouble figuring how to use set user or group ID on execution. I soing this: cp /usr/bin/id id ; chown irc:irc id ; chmod ug+s id ; then as a regular user, I run that id, but the user never switches to the file owner... what I'm doing wrong? 1211990900 M * franck34 Bertl: qmail is a shit to install, there is no package. Some installation guide exists. Qmailrocks is one of them. Just follow, compile, etc ... The only diff in a vserver is to use init/style with "plain" 1211990924 M * pmenier franck34: qmail is in dotdeb packages 1211990931 M * rgl franck34, I think thats not true anymore. qmail is now open source. 1211990951 M * franck34 hu??? since when ? 1211990966 M * pmenier deb http://packages.dotdeb.org etch all 1211990970 M * franck34 if i apt-get install qmail + clamav + spamassin + imap etc, all is correct ? 1211990985 M * pmenier yes i run them since several years. It works fine 1211991003 M * rgl franck34, its even better... its public domain. 1211991004 M * franck34 years ?? 1211991007 M * rgl franck34, http://www.qmail.org/not-open-source.html 1211991017 M * pmenier yes... 2/3 don't remeber exactly 1211991033 M * mnemoc qmail and all djb software was released as public domain some months ago 1211991040 M * franck34 2 i think, the date of my last server install 1211991046 M * franck34 it's a good news 1211991061 M * Bertl well, I consider it a desparate move :) 1211991069 M * franck34 arg 1211991081 M * mnemoc http://lwn.net/Articles/260481/ 1211991095 M * rgl any help with me OT? :D 1211991123 M * Bertl rgl: should work, unless you are missing a cap 1211991125 M * franck34 ha no it's nov 30 2007 1211991131 M * franck34 so less that one year 1211991153 M * mnemoc franck34: that's why I said "some months ago" :) 1211991169 M * rgl Bertl, How do I check? 1211991178 M * franck34 mnemoc: yep, pmenier was saying 2/3 years 1211991183 Q * Hurga Remote host closed the connection 1211991190 M * Bertl rgl: first, check that it works on other suid binaries 1211991191 M * franck34 forget probably misunderstood 1211991202 M * rgl Bertl, su works :D 1211991210 M * franck34 so i just waste 4 hours to install all qmailrocks from source ? 1211991211 M * franck34 :) 1211991216 M * Bertl franck34: btw, what's the advantage of qmail compared to e.g. postfix? 1211991232 M * pmenier at the beginning i had to get qmail-src 1211991246 M * pmenier and build-qmail 1211991262 M * Bertl rgl: then compare the flags/attributes of the su binary with that one 1211991293 M * Bertl rgl: also make sure that the user actually can be su-ed to 1211991357 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1211991362 M * rgl Bertl, oh it works... :D 1211991385 M * rgl Bertl, I was confused with this: uid=1000(rgl) gid=1000(rgl) euid=39(irc) egid=39(irc) groups=20(dialout), 1211991399 M * rgl Bertl, it "only" changes the euid 1211991415 M * rgl Bertl, is there a way to drop all the suplementary groups? 1211991572 M * franck34 Bertl: avantage ? i don't know ? why ? because i started with qmail and because a crash i need to reinstall quickly. I'll learn postfix later but i must do it ! 1211991682 M * Bertl rgl: 'man 7 credentials' look for setgroups 1211991697 M * Bertl franck34: fair enough! 1211991724 M * rgl Bertl, you known what package contains that man page? 1211991753 M * rgl Bertl, I known about setgroups(2) though. 1211991767 M * Bertl man-pages-2.64-1mdv2008.0 here :) 1211991786 M * rgl Bertl, but I was thinking in using just shell wizardy. 1211991807 M * rgl oh, so the man pages here (debian etch) are somewhat ancient. 1211991829 M * Bertl who would have guessed :) 1211991855 A * rgl confiscates the evil hat from Bertl 1211991904 M * rgl Bertl, not even lenny has credentials *G* 1211991952 M * Bertl *phew* good that 'my' distro has the necessary 'credentials' :) 1211991992 J * pisco_ ~pisco@tor.noreply.org 1211992018 M * rgl Bertl, found it at http://linux.die.net/man/7/credentials *G* 1211993003 J * hparker ~hparker@linux.homershut.net 1211993023 M * rgl Bertl, so having root run a setuid to some other user is a no no because that setuid binary can switch back to root :/ 1211993091 M * Bertl that's why suid-root apps take special care to drop all priviledges 1211993301 J * mick_work_ ~clamwin@h-74-2-196-226.miatflad.covad.net 1211993301 Q * mick_work Read error: Connection reset by peer 1211993303 N * mick_work_ mick_work 1211994266 Q * pisco_ Ping timeout: 480 seconds 1211994369 Q * pmenier Quit: Konversation terminated! 1211995802 Q * alex_ Remote host closed the connection 1211995910 J * pisco ~pisco@tor.noreply.org 1211996183 J * balbir ~balbir@122.167.176.181 1211997310 Q * samuel Quit: samuel 1211997423 J * larsivi ~larsivi@144.84-48-50.nextgentel.com 1211998414 Q * balbir Ping timeout: 480 seconds 1211999137 J * Solver ~robert@rbrockway.tor.istop.com 1212000822 M * pisco 1ls 1212000923 J * Linus ~Nuhks@bl7-133-47.dsl.telepac.pt 1212001125 Q * duckx Remote host closed the connection 1212001287 J * duckx ~Duck@81.57.39.234 1212001306 Q * pisco Ping timeout: 480 seconds 1212001949 Q * bfremon Quit: Leaving. 1212002184 Q * dna Ping timeout: 480 seconds 1212002206 J * bfremon ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1212002959 J * larsivi_ ~larsivi@144.84-48-50.nextgentel.com 1212003050 Q * larsivi Ping timeout: 480 seconds 1212003217 Q * bfremon Remote host closed the connection 1212003422 J * bfremon ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1212005368 J * xdr ~xdr@gote2.233.cust.blixtvik.net 1212006291 Q * bfremon Quit: Leaving. 1212006489 J * bfremon ~ben@lns-bzn-29-82-248-242-139.adsl.proxad.net 1212007273 J * yarihm ~yarihm@84-74-147-84.dclient.hispeed.ch 1212007307 J * derjohn_mob ~aj@e180206217.adsl.alicedsl.de 1212007587 J * marv ~marv@74.57.218.127 1212007589 N * marv marv_ 1212007659 M * marv_ SET_RLIMIT goes in /etc/vservers//cccapabilities and QUOTA_CTL goes in /etc/vservers/ --set --ccap SET_RLIMIT 1212009446 Q * bfremon Quit: Leaving. 1212010088 Q * bonbons Quit: Leaving 1212011268 N * DoberMann DoberMann[PullA] 1212011583 M * marv_ VM: killing process ircd(20806:#40041) 1212011589 M * marv_ why am i always getting this? 1212011600 M * daniel_hozac because you're running out of memory. 1212011611 N * DoberMann[PullA] DoberMann[ZZZzzz] 1212011611 M * marv_ 2gb free 1212011622 M * marv_ 200some megs free on the guest 1212011639 M * Bertl check /proc/virtual//limits 1212011649 M * marv_ on guest on host? 1212011658 M * marv_ host 1212011755 M * marv_ VM = Virtual Memory? 1212011774 M * Bertl check for the last column 1212011792 M * Bertl (maybe upload it to paste.linux-vserver.org) 1212011867 M * marv_ http://paste.linux-vserver.org/12123 1212011891 M * Bertl RSS: 15897 0/ 196608 65536/ 196608 62 1212011904 M * Bertl that makes 62 hits on the RSS limit 1212011921 M * Bertl roughly 62 kills since running :) 1212011933 M * marv_ the box has 191megs free 1212011937 M * marv_ the guest 1212011943 M * marv_ + whats cached 1212011948 M * daniel_hozac right now, yes. 1212011960 M * daniel_hozac presumably that ircd process consumed quite a bit of RAM. 1212011973 M * Bertl then got killed, all free again 1212011978 M * marv_ other servers with LESS ram can run it 1212011979 M * Bertl sounds like a memory leak 1212011984 M * marv_ on the same host 1212011998 M * marv_ ? 1212012019 M * Bertl different version, different config, different usage pattern 1212012026 M * marv_ same version, same compile 1212012030 M * marv_ same OS on guest 1212012035 M * marv_ same compilers/version on everything 1212012046 M * marv_ compiles/runs fine on every guest execpt this one 1212012151 M * Bertl well, keep watching the limit 1212012164 M * Bertl look when it gets high who uses up the memory 1212012233 M * marv_ the total ram on the host isn't the issue i have 2gb free... 1212012238 M * marv_ i removed the rss limits and it loads 1212012249 M * marv_ its impossible that ircd uses over 100megs of ram 1212012266 M * Bertl well, check again with proc 1212012277 M * Bertl the same table will tell you the maximum used so far 1212012343 M * marv_ , 1212012345 M * marv_ m'k thx 1212012356 M * marv_ maby it just needs a hole lot durning the _load_ 1212012590 M * marv_ i upped the rss.hard limit... works now 1212012600 M * marv_ i dont understand how it can hit the limit thow 1212012606 M * marv_ hehe neways that was the issue 1212013657 J * Satriani ~Nuhks@bl7-134-142.dsl.telepac.pt 1212014023 Q * Linus Ping timeout: 480 seconds 1212015788 Q * lagann Quit: Leaving 1212015910 J * lagann ~distro@c-66-30-110-51.hsd1.ma.comcast.net 1212016311 Q * dowdle Remote host closed the connection 1212016625 Q * lagann Quit: Leaving 1212016658 J * lagann ~distro@c-66-30-110-51.hsd1.ma.comcast.net 1212016913 J * geb ~geb@41.207.17.52 1212016921 M * geb hi 1212016961 Q * brag_ Remote host closed the connection 1212017064 M * Bertl hi 1212017103 M * geb howdy ? 1212017181 Q * lagann Quit: Leaving 1212017296 J * lagann ~distro@c-66-30-110-51.hsd1.ma.comcast.net 1212017543 Q * geb Quit: Quitte 1212017771 Q * MatBoy Remote host closed the connection 1212017837 Q * Satriani Remote host closed the connection 1212018431 J * brag ~bragon@2001:7a8:aa58::1 1212018896 Q * yarihm Quit: This computer has gone to sleep 1212019194 Q * virtuoso Ping timeout: 480 seconds